You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/running/jobreport.md
+6Lines changed: 6 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -133,6 +133,12 @@ GPU Specific Values
133
133
The `jobreport` tool requires the application to complete successfully.
134
134
If the application crashes or the job is killed by `slurm` prematurely, `jobreport` will not be able to write any output.
135
135
136
+
!!! warning "Too many GPUs reported by `jobreport`"
137
+
If the job reporting utility reports more GPUs than you expect from the number of nodes requested by SLURM, you may be missing options to set the visible devices correctly for your job.
138
+
See the [GH200 SLURM documentation][ref-slurm-gh200] for examples on how to expose GPUs correctly in your job.
139
+
When oversubscribing ranks to GPUs, the utility will always report too many GPUs.
140
+
The utility does not combine data for the same GPU from different ranks.
141
+
136
142
!!! warning "workaround known issue on macOS"
137
143
Currently, there is an issue when generating the report file via `jobreport print` on the macOS terminal:
0 commit comments